Implementación de un sistema de registro de accesos para empleados mediante reconocimiento facial con aprendizaje profundo en una bitácora digital para una empresa de seguridad.

Abstract

The present study proposes the development of an access registration system for employees using facial recognition with deep learning, whose objective is the efficient recognition of the face of employees of the security company using deep learning algorithms, thus improving the facial recognition system. to avoid false positives and false negatives. The research used an applied and specific methodology, as well as a SCRUM framework, using facial recognition techniques, convolutional neural network (CNN), deep learning algorithms to optimize times in the entry and exit process of the company's employees. company. A functional prototype was designed and developed that was adapted to the improvements that were announced as the meetings were held, both virtual and in-person, a series of tests were developed in controlled environments with each version of the system from the previous version. 1.0 to version 4.0 to check the correct operation in each of them of the connection with the database, the loading of the data to the database, the display of the person's data once their face is recognized to indicate your entry or exit, the continuous recognition of the face in the facial recognition section, the relevant validations and the facial recognition system with deep learning, the final version of the system is version 4.0, this version was implemented correctly in the company, meeting the expectations of the company and functioning optimally as requested.